Summary
Asia-Pacific Network Operations and Management Symposium
2022
Session Number:PS2
Session:
Number:PS2-02
Study of Information Identification Technology Considering of Overlapping and Synonyms Words
Haruka Osanai, Masaru Sakai, Sho Kanemaru, Kensuke Takahashi, Satoshi Kondou,
pp.-
Publication Date:2022/09/28
Online ISSN:2188-5079
PDF download
Summary:
When multiple companies in different industries collaborate to provide ICT services, the systems used by each company differ and the format in which information is expressed differs from system to system. Therefore, information cannot be distributed without manual information conversion. To promote provision of services through the cooperation of multiple companies, the distribution of information among them needs to be automated. In this case, information content needs to be automatically identified to determine the distribution destination. In this paper, we propose a method that can automatically identify information by using a combination of distributed representations and a synonym conversion. The evaluation results show that the method can identify information with an existing type with a 100% accuracy, but needs to be improved regarding identification of whether information is a new type.